Task background training increased mean retrospective detection area in eight SEEG streams under one-second waits and budgets of 0–6 background commands per eligible minute
We tested two fixed readers in eight public SEEG recordings from OpenNeuro ds004473 version 1.0.2. We asked whether training with task background gave better detection before a keypress at the same maximum background command budget. We reused fixed scores and later test blocks from an earlier study. We tested every distinct score threshold with a one-second waiting rule. Test labels selected the best threshold at each budget. We averaged detection over budgets from 0 to 6 commands per eligible minute. Mean detection area divided by six was 13.32% with rest training and 16.47% with task background training. The paired mean change was +3.14 percentage points. A descriptive 95% patient bootstrap interval was +0.54 to +5.88 points. Seven of eight patients had a positive change. A strictly increasing score change preserves this complete frontier. Thus a score offset alone cannot explain the difference. Earlier selected thresholds did not keep every later command rate within its budget. The frontier is an observed limit for fixed readers. It does not show live control or stimulation effects. The package contains data, code, proofs, figures, complete threshold tables, source records, and repeat instructions.
SEEG · brain-computer interfaces · threshold selection · command waiting rule · detection frontier · retrospective evaluation
References
4- usesA higher score threshold changed exact command times in eight fixed SEEG streams with a one-second waiting period sci:sha256:f3d4cb75d8fa2ee27205d550a7dff8e125e5fafcc430399c663e4072ed9a50fc
We use its earlier result on changed command times under a waiting rule as background for the complete threshold method.
- usesTask background training increased mean detections and background commands in eight later SEEG streams at an early target of one command per eligible minute sci:sha256:d193eb1dd45105cd24c75e54cf8edf1698b27eda17d7a057b9662c28f33a06be
We use its unchanged data, fitted readers, scores, labels, masks, and source records. Complete later command frontiers resolve its stated limit about comparison under the same later budget.
- supersedesTask background training increased mean retrospective detection area in eight SEEG streams under one-second waits and budgets of 0–6 background commands per eligible minute sci:sha256:26e6f4f7757532e568148a1013ccc18ff5e1762f5cc226aa32d701c1b6b3c512
This version makes the package check ignore metadata added by sci fetch. Data, results, figures, and the scientific report are unchanged.
- usesdoi:10.18112/openneuro.ds004473.v1.0.2
The analysis reuses the fixed SEEG recordings and public source metadata from OpenNeuro ds004473 version 1.0.2 through the preserved parent publication.
Cited by
3 publicationsPublic catalog entries on Sciscape.
- usesA 0.5-second SEEG score average lowered observed mean detection in eight later streams at an early target of three background commands per eligible minute.
This study uses the actual scheduled background-count definition and the retrospective maximum-budget frontier. It keeps all partitions because counts can rise at a higher threshold.
- usesPost-response training gave no clear mean phase AUC gain after 32 fit trials in eight offline SEEG readers
Its task background comparison and one-second response guards informed the question. Our new test examines short post-response windows inside that omitted region. We did not reuse its fitted models or stream features.
- usesThree consecutive SEEG scores gave no clear mean gain in retrospective detection area at maximum budgets of 0 to 6 background commands per eligible minute in eight patients
This publication supplied the complete threshold frontier definition and the simultaneous threshold replay method. The new study uses these methods to compare one score with three consecutive scores after a fixed reader. It checks every threshold with a separate scalar replay.